[3H]cAMP competition binding assay (Nordstedt et al., 1990) was implemented in our laboratory by Vonk et al. (2008, 2009). This assay served as standard for cAMP detection and is still in use today, as a reference method at the beginning of projects with new GPCRs or new recombinant cell lines. However [3H]cAMP competition binding assay is an in vitro endpoint assay only allo-wing detection of the analyte in cell lysates.
The use of cAMP biosensors provides significant advantages for monitoring of cAMP production in real time without the need for cell lysis. Due to suitable affinity for cAMP (~ 1 µM), the monomolecular structure of the biosensor construct and its uniform distribution throughout the cytosol Epac2-camps bio-sensor (Nikolaev et al., 2004) we selected for detection of cAMP upon GPCR stimulation.
For biosensor expression we first used common transfection reagents Lipofectamine 2000 (Invitrogen Life Technologies) and ExGen 500 (Fer-mentas). Inconsistent expression levels between independent assays posed problems for our assay development. Increases in the amount of DNA and lipophilic reagent caused increased cytotoxicity but did not result in higher protein expression. Also to perform dose-response experiments with a reason-able amount of data points we needed high number of cells expressing the bio-sensor. It was clear that the use of transfection reagents poorly suits the development of a high-throughput screening assay due to their relatively high cost and the constant need for large amounts of purified plasmid DNA.
To overcome these limitations we implemented the BacMam technology.
Epac2-camps gene under the human cytomegalovirus promotor was cloned into the baculovirus vector and the polyhedrin promoter sequence driving protein expression in insect cells was eliminated. After obtaining the virus the condi-tions for cell transduction were optimized for the B16F10 cell line expressing the melanocortin MC1 receptor. We found that suitable sensor expression levels were achieved at multiplicity of infection of 200 to 400 and by incubation of the mammalian cells with baculovirus for 2 to 4 h. To enhance protein expression after the transduction step, sodium butyrate, a histone deacetylase inhibitor was used at 10 mM. Sufficient protein expression for functional assays was achieved 24 h post-transduction and remained such for a further 24 h. It was de-monstrated that the assay was flexible and had a convenient time frame between the transduction of cells and the functional measurements themselves (Paper I).
To verify expression and functionality of the MC1 receptor in B16F10 the cells were stimulated with MC1R endogenous agonist melanocyte stimulating hormone α-MSH. The accumulated cAMP was measured using the [3H]cAMP competition binding assay (Figure 1). Upon 100 nM α-MSH a twofold increase in cAMP level was detected. Receptor activation upon stimulation with α-MSH alone and in presence of 100 µM IBMX (phosphodiesterase inhibitor) was much higher (ninefold increase over the basal cAMP level). These results showed that B16F10 cells are suitable for MC1R activation studies.
Firgure 1: Influence of a-MSH and IBMX on cAMP levels in B16F10 cells. Cells were treated with the ligands for 20 min at 37 °C. The accumulated cAMP was detected using the [3H]cAMP competition binding assay. Graph showing data from a represen-tative experiment carried out in sextuplicates (n=2).
After verification of MC1R expression and functionality in B16F10 cells, we started to develop the cAMP biosensor assay. Commonly for the cAMP accu-mulation assays, phosphodiesterase inhibitors are used to amplify the signal.
Accounting for phosphodiesterases, enzymes that hydrolyze cAMP to a non-signaling AMP, is significant, because they determine the lifetime of the “effec-tive pool” of cAMP measured in the assay (Willoughby and Cooper, 2008; Hill et al., 2010). Therefore we first investigated the role of inclusion of IBMX in the biosensor assay. It was determined that 100 µM IBMX did not affect EC50 values for α-MSH or forskolin (a direct activator of adenylate cyclase) but the inhibitor did also not improve the biosensor based detection system. IBMX decreased the dynamic range of the assay by elevating the basal level of cAMP (the change in FRET signal was smaller, compared to cells assayed without addition of phosphodiesterase inhibitor). Compared to the in vitro assay, where the accumulated cAMP is measured after cell lysis, the biosensor responds to immediate changes in cAMP. With the use of the biosensor it was now possible to detect changes in cAMP concentrations in real time, without the need for additional signal enhancement.
Bivalent metal cations are known to play an important role in ligand binding to GPCRs but the exact mechanisms still remain elusive and vary depending on the ion and on the receptor in question. It has been shown that Zn2+ and Ca2+
ions modulate binding of ligands to the MC1R (Holst et al., 2002; Kopanchuk et
al., 2005). To determine the effect of bivalent cations on receptor activation in B16F10 cells, cells were first treated with EGTA or EDTA, washed with phosphate buffered saline and thereafter, fixed concentrations of Ca2+ or Mg2+
were added to the reaction mixture. Upon removal of bivalent cations with 1 mM EDTA, the potency of the MC1R agonists (pEC50 = 6.8 ± 0.6 for α-MSH) was significantly lower than in presence of bivalent cations (Figure 2). Addition of 1 mM Ca2+ was found to be necessary for high-potency agonist effect (pEC50 = 9.2 ± 0.4), whereas 1 mM Mg2+ also increased the potency of agonists but to a lesser extent (pEC50 = 7.9 ± 0.1). Similar effects were observed for all MC1R agonist tested (Paper I). After treatment with EGTA, which binds Ca2+
with higher affinity than Mg2+, the obtained results were comparable to data observed after EDTA treatment. These results together demonstrated that the pEC50 values strongly depend on the ionic composition of the assay buffer. The effect of the bivalent cations must be caused at the step of ligand binding to receptor, because activation of cAMP signaling cascade is not abolished by their removal. In this light, all of the following experiments we carried out in phosphate buffered saline in the presence of 1 mM Ca2+, if not stated otherwise.
Figure 2: The effect of Ca2+ and Mg2+ on MC1R activation by α-MSH. B16F10 murine melanoma cells were transduced with BacMam-Epac2-camps virus for 3 h and further incubated for 21 h in complete growth medium supplemented with 10 mM sodium butyrate. Cells were washed with DPBS containing 1 mM EDTA prior to experiment. Chelating agents were removed and cells were assayed in DPBS (with or without 1 mM Ca2+ or Mg2+) upon 10 min treatment with a full agonist, α-MSH. The maximal FRET change was normalized to 100 % response. Without Ca2+ and Mg2+: pEC50 = 6.8 ± 0.6; in presence of 1 mM Mg2+: pEC50 = 7.9 ± 0.1; in presence of 1 mM Ca2+: pEC50 = 9.2 ± 0.4; (n=4). Graph showing data from a representative experiment.
The MC1receptor activation was further characterized by a set of known MC1R ligands (Figure 3). In B16F10 cells all studied MC1R agonists caused a con-centration dependent increase in cAMP concon-centration. α-MSH, β-MSH, NDP-α-MSH behaved as full agonists for MC1R with similar sensor activation profiles, while their activation level remained 75 % of the level achieved by forskolin.
Partial agonists (SHU-9119, MS-05, HS-024) achieved a level of 75% of full agonist activation (Paper I). The selective low molecular weight MC4R agonist I-THIQ had no effect on MC1R activation in our system. These data are in agreement with previously published efficacies for these ligands. It was demonstrated that the cAMP biosensor assay is able to distinguish between full and partial agonists, which is important for characterization of newly synthe-sized ligands. In vitro cAMP accumulation assays and reporter gene assays are sometimes insensitive to detection of partial agonism, because of the signal amplification along the signaling cascade or due to measurements of the amplified readout.
Figure 3: Influence of different concentrations of MC1R ligands on cAMP bio-sensor signal in B16F10 cells. Cells were transduced with BacMam-Epac2-camps virus for 3 h and further incubated for 21 h in complete growth medium supplemented with 10 mM sodium butyrate. Cells were treated with ligands for 10 min at 37 ºC. Measure-ments were performed in DPBS + 1 mM Ca2+ for full receptor activation. The maximal FRET change was normalized to 100 % of the full agonist NDP-α-MSH response.
pEC50 ± S.E.M. (n=4): NDP-α-MSH: 9.9 ± 0.4; α-MSH: 9.2 ± 0.4; β-MSH: 8.74 ± 0.06;
SHU-9119: 8.9 ± 0.3; MS-05: 7.96 ± 0.03; HS-024: 7.27 ± 0.13; I-THIQ showed no activation (n=2). Graph showing data from a representative experiment.
The study on MC1R in B16F10 cells demonstrated that the developed BacMam system for cAMP biosensor expression is suitable for GPCR activation studies and the calculated Z´-factor values > 0.6 for all MC1R specific agonists and forskolin confirmed its suitability for high throughput screening (Paper I).
In 2011 a third generation TEpacVV biosensor was constructed and characte-rized by Klarenbeek et al. The sensor consists of a part of Epac1 protein fused between a bright fluorescent protein mTurquoise and a tandem acceptor consisting of two Venus proteins. The new sensor was shown to have better signal-to-noise ratio and dynamic range compared to previous versions of Epac based cAMP biosensors (Klarenbeek et al., 2011).
The group kindly provided us with the new cDNA construct that we cloned into the BacMam expression vector. The BacMam viruses containing the
TEpacVV biosensor gene were generated analogues to the work with Epac2-camp biosensor. After generation and optimization of the expression system, TEpacVV biosensor was compared the former, Epac2-camp biosensor. HEK293 cells stably expressing dopamine D1 receptor were transduced with BacMam viruses containing the gene of either of the biosensors and the cells were assayed on the next day. The dynamic range (FRET change window) was almost two times wider for the TEpacVV biosensor, but no differences in the potencies were deter-mined from the dose-response curves obtained with two different biosensors.
The calculated EC50 values were approximately 1µM for forskolin and 1 nM for dopamine (Figure 4). The small difference in the affinities of the cAMP-binding domains of Epac1 and Epac2 proteins used in the biosensors (Kd ~ 4 and 1.2 µM, respectively; De Rooij et al., 2000) did not show discrepancies in the measured potencies for the known compounds. These data confirmed that switching from Epac2-based sensor to the Epac1-based biosensor for higher sensitivity in terms of FRET span is safe and justified. From then on TEpacVV biosensor was used in all of the following studies.
Figure 4: Cellular responses to forskolin and dopamine measured with Epac2-camps and TEpacVV biosensors. HEK293 cells stably expressing dopamine D1 recep-tor were transduced with BacMam virus for 3 h and further incubated for 21 h in complete growth medium supplemented with 10 mM sodium butyrate. Cells were treated with adenylate cyclase activator forskolin or D1R agonist dopamine for 10 min at 37 ºC. Graph showing data from a representative experiment (n=2) performed in triplicates.
Dopamine has several functions in central nervous system including voluntary movement, feeding, affect, reward, sleep, attention, working memory, and learning (Beaulieu and Gainetdinov, 2011). Dopamine receptors are thus targets for many drugs. Although many dopaminergic ligands have been on the market for a long time (agonists mainly in the treatment of Parkinson’s disease and antagonists as antipsychotics in the treatment of schizophrenia), development of more selective and potent dopaminergic compounds is still of great importance (Reinart-Okugbeni, 2012; Beaulieu et al., 2015). It is therefore important to have a functional assay for characterization of potencies of dopaminergic li-gands in parallel to ligand binding experiments. All five subtypes of dopamine receptors signal via the adenylate cyclase pathway. D1 and D5 receptors are coupled to the Gs family proteins, hence agonist binding activates cAMP synthesis. D2, D3 and D4 receptors are coupled to the Gi family proteins and upon receptor activation cAMP synthesis is inhibited. We had successfully used the cAMP biosensor to monitor Gs coupled receptor activation (MC1 and D1 receptors), but now the assay system needed to be adapted for measurements of decreases in cAMP levels induced by Gi coupled receptor activation.
To monitor the decrease in cellular cAMP, its cellular level is increased by pretreatment with a direct AC activator forskolin. The assay conditions were optimized by varying the concentrations of forskolin between 1 and 50 µM.
Upon cell stimulation with 10 µM forskolin together with receptor agonist, the dose dependent inhibition of cAMP production was detected in the widest dynamic range (15–20% of FRET change). Gi coupled receptor agonist potency in this type of assay is described by the IC50 value (Figure 5; Paper II).
Figure 5: Inhibition and activation of cAMP synthesis. HEK293 cells stably expres-sing dopamine D1 or D3 receptor were transduced with BacMam virus for 3 h and further incubated for 21 h in complete growth medium supplemented with 10 mM sodium butyrate. HEK293-D1R cells were treated with serial dilutions of dopamine and the response was measured 10 minutes after agonist treatment (blue). For D3R stimu-lation cells were treated with serial dilutions of dopamine together with 10 µM forskolin (direct activator of adenylate cyclase). The response was measured 10 minutes after agonist treatment (green). Graph showing data from a representative experiment (n=3).
Our general aim is to approach and study GPCR systems from different direc-tions. In Paper II we have introduced an approach to gather and integrate expe-rimental data from ligand binding experiments, G protein activation studies and measurements of the second messenger levels. The strategy relies on the two basic properties, experiments are being performed using baculovirus constructs and the detected signals are based on molecular fluorescence. In the future, we aim to combine different experimental readouts into new meaningful infor-mation about the signal transduction mechanism of GPCRs.
Characterization of the biological activity of novel compounds requires pre-cision and good assay reproducibility. We have provided a step-by-step protocol for generation and application of the BacMam based cAMP biosensor assay (Paper III). We also present a thorough description of a novel protocol for determination of virus titer with a cell size-based assay adapted and modified in our laboratory. To help other groups trying to set up the BacMam based expression system tips and observations have been included as notes at the end of the paper. This publication is a tool to disseminate our cumulative knowledge and observations in the field on cAMP detection and GPCR ligand characte-rization.
In addition to the functional characterization of low-molecule weight com-pounds and peptide ligands, the developed cAMP biosensor based assay system could also be applicable for studies of hormone receptors and their ligands, glycoproteins (Paper IV). The LH receptor is known to bind two proteins of the gonadotropin family, human luteinizing hormone (hLH) and human chorionic gonadotropin (hCG). Despite the high level of structural homology and action at a shared receptor, hLH and hCG have vital and unique roles in human develop-ment and reproduction. Our objective was to examine the differences or simila-rities of LH receptor activation by hLH and hCG. No significant differences in receptor activation and formation of cAMP in COS-7 cells expressing the re-combinant LH receptor (Müller et al. 2003) were determined in this study during the first 120 minutes from the addition of hormones (Paper IV). The specificity of gonadotropin action at the LH receptor must be achieved by another mechanism or as a result of differences in the long-term signaling (Casarini et al. 2012; Choi and Smitz 2014).
Next, the biological activity of hCG preparations of different origins were compared. A four times lower potency for urinary hCG (EC50=100 pM) compared to the recombinant hCG (EC50=25 pM) was determined (Figure 6;
Paper IV). The difference between the measured potencies may be explained by the different manufacturing and the subsequent stability of the urinary purified hormone compared to the recombinant preparation. Also, different patterns of glycosylation of recombinant and urinary preparations may influence the detected biological activity of the hormone. Today the produced hormone preparations are calibrated against the international reference preparations using immunoassays. These assays are based on epitope detection and generally do not account for the proper folding of the protein. The cAMP detection based assay system could be an important addition for quantification of active
amounts/concentrations of the hormones in pharmaceutical preparations, where the information about the fraction of biologically active form of the protein is of highest importance.
Since for evaluation of the biological activity of hormones LH receptor acti-vation was used as the mediator of the biological signal (a natural biosensor), it opened a possibility to use the developed system for quantification of active concentrations of hCG, which is of special interest in the field of in vitro fertilization.
Figure 6: Influence of different concentrations of hCG preparations on cAMP bio-sensor signal. COS-7 cells stably expressing LH receptor were transduced with BacMam-TEpacVV virus for 3 h and further incubated for 21 h in complete growth medium supplemented with 10 mM sodium butyrate. Cells were treated with ligands for 30 min at 37 ºC. For urinary hCG (green): EC50=100 pM; for recombinant hCG (blue):
EC50=25 pM. Graph showing data from a representative experiment (n=4) performed in triplicates.
Human chorionic gonadotropin is produced by trophoblasts, the cells that surround the growing human embryo and later form the placenta. Different isoforms of hCG (primarily hyperglycosylated hCG and the β-subunit of hCG) have been shown to be useful as biomarkers for embryo selection to improve the pregnancy chances after embryo transfer (Ramu et al. 2011; Butler et al.
2013). Today the transplanted embryos are selected based on their morphology and physicians’ expertise and experience using grading systems based on the stage and morphological appearance of the embryo, with the success rate of implantation still under 50%. The high sensitivity (LOD: 5 pM or 4 mIU/mL) determined from a calibration curve using recombinant hCG, suggests the possibility to use cAMP biosensor assay for detection of hCG as a biomarker in the analysis of embryo spent culture media (Paper IV). The LODs of immunoassays used in previous quantitative studies are in the low mIU/mL range (Butler et al. 2013; Stenman et al. 2013; Xiao-Yan et al. 2013). However,
since in the cAMP biosensor assay only the active portion of the total hormone concentration is eliciting a measurable response, the possibly higher total hormone concentration of the probe (detectable by immunoassays) may result in some trade-off between signal specificity and sensitivity. Addressing the applicability of the developed BacMam based cAMP biosensor assay for challenging tasks as unbiased embryo selection is subject for future studies.